Explaining higher education social sciences students’ misuse of generative artificial intelligence: evidence from a multidimensional ethics scale

Abstract

The growing presence of generative artificial intelligence (GenAI), such as ChatGPT, in higher education raises significant ethical concerns, particularly regarding its misuse in academic essay writing. This study examines how university students enrolled in social science programs at two Spanish universities ethically evaluate a typical misuse of GenAI—namely, producing academic essays with minimal human editing. Drawing on the Multidimensional Ethics Scale (MES), four ethical dimensions—moral equity, relativism, consequentialism, and deontology—together with gender and employment status, are analyzed using a dual-method approach combining partial least squares structural equation modeling (PLS-SEM) and fuzzy-set qualitative comparative analysis (fsQCA). The PLS-SEM results show that only consequentialism and relativism significantly influence students’ intention to use GenAI, indicating a predominant reliance on perceived utility and contextual social norms. In contrast, moral equity and deontology do not exhibit statistically significant effects. Complementing these findings, the fsQCA identifies multiple, asymmetric causal configurations leading to both acceptance and rejection of GenAI misuse, underscoring that adoption and non-adoption are driven by distinct underlying mechanisms. By integrating correlational and configurational perspectives, this study advances understanding of the ethical complexity surrounding GenAI use in higher education. The findings highlight the need for differentiated educational strategies that account for heterogeneous student profiles and diverse moral reasoning frameworks.

Ähnliche Arbeiten